It doesn't matter if it sounds right to us its what they requested. I'm not sure why you are even asking this forum about this. 

If a pre master had an rms value of -20 but had peaks close to zero that is not such a bad situation as the peaks can be limited to values less than that eg -6 or so then the total rms value of the master can come up accordingly and in stages too. What is more common are pre masters to be sitting around -14 and peaks coming up to -3 dB or so. This results in transients that are still 11 dB above rms values and will still sound pretty good.
 
This would result in a RMS of -13, a good CD volume.....
 
This would be nice of course but.. No, sorry in the real world most clients would consider this to be way too soft. More like -6 ot -7 dB rms for final masters to be nice and loud. You need a great limiter to do this and still maintain transients.

konradh
Craig, That is interesting.  I have not sent anyone tracks yet and was just inquiring, but Sweetwater specifically asked for 16-bit.
 
They also asked for the songs in CD order, which seemed odd to me.  I may rearrange after mastering.

There is a difference between "glass mastering" and bogus mastering. If you are just asking them to glass master then someone has to take the responsibility of preparing 16 bit files.
 
Perhaps there is a mis communication about what sort of mastering you and they are speaking about?
 
best regards,
mike
 
 

Anderton

drewfx1
I'm always baffled by this "leave room" argument.
 
Do mastering engineers not request/use/understand floating point? Or do they, and this is supposed to be some sort of social engineering?

Most people who mix don't test for (or are even necessarily aware of) inter-sample distortion.

 
Most people who are aware of intersample peak distortion do not seem to consider the implication of what the term intersample means.
 
For example; What does half a sample at a 44.1kHz sample rate sound like?
 
A person can demonstrate for them selves that a full 0dBFS clip has to extend for several samples before it is discernible. That is why so many commercial releases actually have 0dBFS clips that go unnoticed without specific analysis.
 
The idea that intersample distortion is problematic is an idea that disregards the idea that no one is capable of noticing actual, real life intersample peak distortion at 44.1kHz. Can someone hear an intersample peak in a 22.05kHz mp3? Maybe. Can someone hear an intersample peak in a 11.025kHz teleconferance call? Probably.
 
In other words, if someone thinks they can hear "intersample peaking" at 44.1kHz what they are probably hearing is a squared wave, 0dBFS clip, with a duration of 10 or more samples. They are certainly not really hearing the actual real life intersample peak that may, or may not have occurred within the very first sample interval of the sequence of samples that would have extended beyond 0dBFS.
 
regards,
mike
 
 
 
spelling

bitflipper

drewfx1
I'm always baffled by this "leave room" argument.
 
Do mastering engineers not request/use/understand floating point? Or do they, and this is supposed to be some sort of social engineering?

Quite right! If they just requested 32-bit files then levels wouldn't really matter. But while some will gladly accept any format you want to send them, others will be specific, perhaps asking for 24-bit files.

Let us say that the "mastering engineer" wanted 24bit integer files. What is the harm in supplying a mix that peaks at -1dBFS?
 
I've heard all the explanations about why the mix should be peaking at *insert number here*dBFS and I'm still waiting to encounter one that actually makes sense from either a technical or aesthetic standpoint.

I suggest you turn your monitoring up if you tend to mix too hot. DAC volume knob is handy.
It is also a good idea to throw a limiter on or a basic master chain on the master bus to check your mix but remove it when printing the mixdown.
16bit does sound dodgy but I suspect it is a safety to prevent any dithering before mastering.
Many Mastering houses will up sample all the tracks and it is not uncommon to clip the AD converters via their analogue chain.
3 overs usually equals a clip and with some gear manufacturers who don't want their products sounding "clipped" too quickly with operator error might even lower the real 0dbfs.
This margin, especially 3 sample overs equaling a clip; it actually might not be actually clipping especially with expensive ADDA Converters. I think it is called "riding the skids"
Yep some go off the rails. LOLOL
 
For us mix guys you just make your mixes as best as you can taking into consideration the intended audience and media. 
 
It is usually up to the mastering house to assemble the order per your instructions. So you got to live with your mixes and try out the order or give it to the producer to decide.
-3 is hot 
 
 
 

mike_mccue

konradh
Craig, That is interesting.  I have not sent anyone tracks yet and was just inquiring, but Sweetwater specifically asked for 16-bit.
 
They also asked for the songs in CD order, which seemed odd to me.  I may rearrange after mastering.

There is a difference between "glass mastering" and bogus mastering. If you are just asking them to glass master then someone has to take the responsibility of preparing 16 bit files.
 
Perhaps there is a mis communication about what sort of mastering you and they are speaking about?
 
best regards,
mike
 
 

I would re-question this as well to Sweetwater. Although some Mastering houses Mastered 16-bit recordings in the 90's, it sounds like Sweetwater is assuming you have your final mix ready and are wanting to go to CD Glass Master, which in the 70's and 80's would technically mean, "Mastering". Make sure you know what they are going to do, because a true Mastering session isn't typically a hand-off from Mixing Engineer to Mastering Engineer. There's reasonable back-and-forth between the two until a solid relationship/trust is formed over time. For making a glass-master, it really is a simple hand-off at 16-Bit 44.1Khz.
 
Definitely get clarity. I messed up on an LP because the press didn't clearly define "High-Volume", which I interpreted as mass-production vs. actual high-volume. Surrounding text indicated the former. I should have gotten better clarity around this. It was a $270 mistake on my part.
 
Don't regret. Get.

Great discussion.
 
John, What the heck did this mean:
"It doesn't matter if it sounds right to us its what they requested. I'm not sure why you are even asking this forum about this."
 
I am asking this forum obviously because I wanted opinions and advice.  I got a great deal of both which has enabled me to select the right mastering service and to prepare my mixes.  This is a Sonar foum and I am mixing in Sonar.
 
Based on the comments, I think I will use a different mastering service.

It means Konrad that you have spoken to the mastering people and they have told you what they require. To me that is the end of it. What this forum thinks is irrelevant.  I doubt you are going to deliver something that is not within the the -3 dB request. 

OK, John, gotcha.  But while I would deliver the format and levels and bit-depth the service requested, based on the feedback here, I will probably select a different service.  Take care.